Hasura GraphQL Engine 2.0

Caching is hugely important and really needs to be integrated at the graph level, but they lock it behind their Cloud offering. Not open source, nor testable in any way before committing.

To further compound, their cloud offering is one of the least compelling I’ve seen. Totally offensive limits on the free tier, a fake $99 tier that is severely limited, and then goes straight to enterprise.

The other day I learned about fly.io and wow, comparing the two it’s just a totally different world. Almost 100x more generous free tier, way more features, and flexible. I have no relation to them, and I do use Hasura for our company, I just hope Hasura steps it up and competes a lot harder.

They shouldn’t lock basically essential features into a cloud-only, closed source model, and they should really push the cloud stuff to work at the edge and honestly just don’t lock people into your cloud (sell a license to use caching how you want and I’d buy it). To be honest I don’t see how they ever fully compete with something like fly which really can focus on infra. I do hope they figure it out.

I thought 1.4 (turns out they called it 2.0) was going to be all about multi-database access, but reading this post I am exceptionally pleased to discover that this release includes a new REST interface feature. And like everything Hasura seems to do, this feature is most elegant. It looks like you can create a REST endpoint for a GraphQL query that you can craft in GraphiQL, and this works for both queries and mutations. This will be so useful for me in backend services, where right now I am using GraphQL, which I don't think really adds value on the backend. I can't wait to try this out.

A thing that makes me happy about Hasura is they seem to have got licensing down right from the beginning, given what we're seeing with projects like Elasticsearch. I look forward to getting Hasura Pro when I can, and I appreciate that the company has set themselves up having customers and making money.